摘要
设计了一类使图像前景缩小的CNN模板,并对模板的鲁棒性进行了研究,结果表明:只要模板参数满足定理中的不等式,CNN就能完成使图像前景缩小的功能;通过实验模拟确认了理论结果在计算机图像处理应用中的有效性.
In this paper, a kind of CNN template is developed to decrease the objects in the images and the robustness of this template is also studied. The results show that the CNN can perform the function of decreasing the objects in the images, giving the template parameters meeting equation of the proposed theorems. The validity of theoretical results in computer image processing applications is confirmed through experimental simulation.
出处
《河南师范大学学报(自然科学版)》
CAS
北大核心
2016年第1期169-173,共5页
Journal of Henan Normal University(Natural Science Edition)
基金
山东省高校智能信息处理与网络安全重点实验室(聊城大学)资助
关键词
细胞神经网络
图像前景缩小
鲁棒性
实验模拟
Cellular Neural Network
image foreground(objeets) deereasing
robustness
experiment simulation